Thoughts regarding auto focus functionality, depth of field

[Robert Camfield]

Movement is one thing; the fast varying direction of traveling birds quite another. With old gear, dealing with motion was primarily a matter of sufficient depth of field/aperture setting, shutter speed, and determining a pre-focusing distance that birds might travel through. With manual lenses, pre-focusing distance of birds in flight can be sheer guesswork.

These days, modern auto-focus capability apparently manages this pretty well, though the time interval available for autofocus technology to determine the focusing distance can be fractions of a second. I've worked up a few details, as follows: Imagine, for example, at a distance of 80 feet, a bird is traveling 25 miles per hour toward the photographer at a 20 degree angle off the longitudinal axis of the lens. This pace of travel constitutes 37 feet per second, though the cosine of the angle translates into an effective travel speed of 34.5 inches per second. If we accept Nikon’s circle of confusion criteria (0.0013 inches), the depth of field for a 600mm at F4.0 at a focusing distance of 80 feet is 17.2 inches. At this distance, much of the depth of field is behind the bird in flight, leaving approximately 4.3 inches DOF in front. Once the shutter is electronically tripped, the camera and lens combination have a mere 0.12 seconds to determine the focusing distance, obtain proper focus by moving the relevant lens elements (perhaps a two elements doublet), and capture the image before the front edge of the bird is outside—i.e., in front of—the front boundary of the depth of field. Nikon—and certainly Canon and Sony also—use stepper and linear motor technologies which apparently provide sufficient focusing speed to handle this. Narrower apertures—or shorter lenses—make for a longer front-side DOF, lowering the focusing-speed constraint.

It seems to me that auto-focus speed is unrelated to pre-release and burst functionality.   

[Ilkka Nissilä]

Movement is one thing; the fast varying direction of traveling birds quite another. With old gear, dealing with motion was primarily a matter of sufficient depth of field/aperture setting, shutter speed, and determining a pre-focusing distance that birds might travel through. With manual lenses, pre-focusing distance of birds in flight can be sheer guesswork.

These days, modern auto-focus capability apparently manages this pretty well, though the time interval available for autofocus technology to determine the focusing distance can be fractions of a second. I've worked up a few details, as follows: Imagine, for example, at a distance of 80 feet, a bird is traveling 25 miles per hour toward the photographer at a 20 degree angle off the longitudinal axis of the lens. This pace of travel constitutes 37 feet per second, though the cosine of the angle translates into an effective travel speed of 34.5 inches per second. If we accept Nikon’s circle of confusion criteria (0.0013 inches), the depth of field for a 600mm at F4.0 at a focusing distance of 80 feet is 17.2 inches. At this distance, much of the depth of field is behind the bird in flight, leaving approximately 4.3 inches DOF in front. Once the shutter is electronically tripped, the camera and lens combination have a mere 0.12 seconds to determine the focusing distance, obtain proper focus by moving the relevant lens elements (perhaps a two elements doublet), and capture the image before the front edge of the bird is outside—i.e., in front of—the front boundary of the depth of field. Nikon—and certainly Canon and Sony also—use stepper and linear motor technologies which apparently provide sufficient focusing speed to handle this. Narrower apertures—or shorter lenses—make for a longer front-side DOF, lowering the focusing-speed constraint.

It seems to me that auto-focus speed is unrelated to pre-release and burst functionality.   

The autofocus system can see when the subject is moving towards or away from the camera and compensate for the processing delay to anticipate where the subject will be at the time of the exposure, but this of course assumes the movement is at least somewhat predictable. For fast moving subjects near the camera the autofocus speed can a limitation as the optical elements need to move greater distances than at longer ranges. I realize that a lot of camera trap shots are precisely taken in this close range as the lack of human presence can help get those close shots.

[paul hofseth]

To elaborate on the usefulness of autofocus, I do not doubt that the camera is able to calculate how many inches it has  to move the focus point in difficult situations.

Not only as an unreformed neoluddist, using manual Nikon, Leitz,  Zeiss, Angenieux  and Kinoptik lenses, but alo as a frequently disappointed autofocus user, I must point out that autofocusssing on the nearest twig and setting the exposure based on the most extreme light situation in the entire frame is NOT  one of the blessings of autofocus, rather it is part of the curse of autofocus (both on my Z7 and my CL devices.

Causa finalis is that the little field that determines where distance and exposure is measured tends to move unpredictably as the camera and its buttonry is handled. If one is lucky the point has stayed in a predictable place and not in any corner.

If, like me, one is more interested in the content of the picture than in all manners of clutter in the viewfinder, (and certainily does not wish to spend time on searching for nd twiddling  pressing any other buttons or controls  than fn1 or 2 for enlarging the central field). That the focus pont is invisibly hidden in a far corner is not easily perceived.

In conclusion, I much prefer taking time to focus and think about depth of field myself  even if birds in flight may escape unrecorded 

p.

[Ilkka Nissilä]

To elaborate on the usefulness of autofocus, I do not doubt that the camera is able to calculate how many inches it has  to move the focus point in difficult situations.

Not only as an unreformed neoluddist, using manual Nikon, Leitz,  Zeiss, Angenieux  and Kinoptik lenses, but alo as a frequently disappointed autofocus user, I must point out that autofocusssing on the nearest twig and setting the exposure based on the most extreme light situation in the entire frame is NOT  one of the blessings of autofocus, rather it is part of the curse of autofocus (both on my Z7 and my CL devices.

Causa finalis is that the little field that determines where distance and exposure is measured tends to move unpredictably as the camera and its buttonry is handled. If one is lucky the point has stayed in a predictable place and not in any corner.

If, like me, one is more interested in the content of the picture than in all manners of clutter in the viewfinder, (and certainily does not wish to spend time on searching for nd twiddling  pressing any other buttons or controls  than fn1 or 2 for enlarging the central field). That the focus pont is invisibly hidden in a far corner is not easily perceived.

In conclusion, I much prefer taking time to focus and think about depth of field myself  even if birds in flight may escape unrecorded 

On recent Nikon mid and higher-end bodies, one can choose quite freely the area within which the AF system can sense subjects and measure focus errors, and finding the subject detected by the camera in a far-edge corner just doesn't happen unintentionally unless the area selected is very large. As for the corners the photographer should always be aware of the image content in the corners, if not then the method of viewing (eyeglasses, viewfinder eyepoint, color of focus point marker) may need to be reconsidered/adjusted.

I personally use custom wide-area or wide area S most of the time, with occasional single point or 9-point dynamic. For my subjects, narrower is usually better, though not a single line which can turn out to have gaps. I rarely use auto area AF with full freedom left to the camera; such situations can exist when the camera is on a gimbal and I'm shooting video with a clear single human subject, then auto-area AF works fine.

I used to do a lot of manual focusing but today I can't realistically maintain focus on a moving subject with pixel-level sharpness on the eye by turning the manual focus ring, and achieving it even on a static subject can be difficult. For wide angle and macro work, as well as video, manual focusing definitely has its place. AF can not really handle a situation where you want it to get two subjects at different distances equally within the depth of field; it's always trying to focus one subject perfectly whereas a human operator can handle this situation using depth of field and suitable lens with manual focusing. But this is not for me all that common a situation.

I believe on most cameras you can turn off the link between metering and focus area, and/or reduce it. Traditional non-matrix/evaluative metering patterns are still available.

[Anthony]

If the focus point is jumping around unpredictably, this may be because the rear LCD is set to touchscreen operation. The nose then becomes an autofocus point selection tool, with often strange results.

Guess how I know this 

[paul hofseth]

Thanks for well meaning advice.

I always disable touch screens, ( I would actually prefer cameras without screens, just viewfinders,  except for menus if they absolutely have to have them rather than building the right machine in the first place(yes, I do realize that people have different needs and that an infinitely adjustable device has a bigger market and can be produced more cheaply)) Autoexposure and auto ISO is fine and to be able to have the pictures on my large computer screen in the evening is excellent, but excessive buttonry and wheeels that cannot be disabled is a pestilence. I have not yet used cyanacrylate on my Z7, but have been tempted. My R9 at least has a button that has to be depressed in order to move a crucial control wheel.

My views of screens and instant picture review comes from learning to take pictures by using a IIIc . Then I had to wait for weeks to have Kodachrome developed , or for myself to mix Microphen or similar Ag-halide treatments  and a new load of fix to get the Ilford FP3 or Kodak Panatomic developed and enlarged. Being able to snap brief moments is the main reason I gave up on using Linhofs even though rise and skewing of the lensboard and the bigger format had its advantages. So by now I am stuck with  portable computers with their multitude of choices to serve my glass library.

I use spot metering because I like whatever I focus on to be correctly exposed, and have pensioned off my Metrawatt for measuring incident light. The various metering patterns are probably excellent but my experience with such in winter snow has reinforced my selective metering urge

The focus+metering point inadvertently does move around when I maneuvre the camera out of my rucksac or even when i grasp the apparatus from the  handstrap and happen to touch the no-joy stick.

Giving autofocus liberty to roam is not my idea of focusssing on a subject especially since I have had several instances of uninteresting background sharper than the main subject (example: twigs intsead of deer)

A corollary of these views has to be that to take photographs may need time. And so be it.

p.

[Robert Camfield]

Ilkka:

Thanks for the clarification regarding the autofocus function...Re: anticipation of direction and focus point compensation for delay. I don't have autofocus lenses but, in retrospect, should perhaps have anticipated the workings of the function as you describe, at least in contemporary cameras. 

Paul: I generally share your perspective, but will surely succumb to the convenience and capability of modern technology before long. On this point, the advantage appears to be managing movement, exclusively. To the extent time is availing...a few seconds or longer...manual focus is sufficient and, arguably, significantly more engaging. I much utilize the depth-of-field preview function.

[Ian Watson]

excessive buttonry and wheeels that cannot be disabled is a pestilence. I have not yet used cyanacrylate on my Z7, but have been tempted.

I was excited when I accidentally discovered that my Zf has a custom setting to lock the focus point. It would solve your problem nicely. Alas, it seems to be a feature of only more recent Nikons. Sorry, Paul, but at least Nikon seems to be listening.

[pluton]

I was excited when I accidentally discovered that my Zf has a custom setting to lock the focus point. It would solve your problem nicely. Alas, it seems to be a feature of only more recent Nikons. Sorry, Paul, but at least Nikon seems to be listening.

I believe locking the focus point is an Expeed 7 innovation, so Z9, Z8, Z6III, Zf, and Z5II.  Along with reliable AF, a feature sadly not available on my Z6II or Z7II.  Upon acquiring the Z6II and Z7II, I expected to use focus-zoom for critical focusing most of the time, but I'm still disappointed that Expeed 6 was as primitive as it was.  Also, I find it difficult to believe that there is no more space on the Expeed 6 processor chip that focus point lock cannot be added via firmware update.

[Fons Baerken]

On more than a few occassions i noticed the locked focus point moved away from its fixed place on the Zf.

[Ilkka Nissilä]

I believe locking the focus point is an Expeed 7 innovation, so Z9, Z8, Z6III, Zf, and Z5II.  Along with reliable AF, a feature sadly not available on my Z6II or Z7II.  Upon acquiring the Z6II and Z7II, I expected to use focus-zoom for critical focusing most of the time, but I'm still disappointed that Expeed 6 was as primitive as it was.  Also, I find it difficult to believe that there is no more space on the Expeed 6 processor chip that focus point lock cannot be added via firmware update.

All the DSLRs that I can remember have a focus point lock as a physical control, and none of them have Expeed 7.

[pluton]

All the DSLRs that I can remember have a focus point lock as a physical control, and none of them have Expeed 7.

Alas, retrofitting a lever to a Z camera is "not in the cards".